Summary: Online Load Balancing \Lambda
Yossi Azar y Andrei Z. Broder y Anna R. Karlin y
The setup for our problem consists of n servers that must complete a set of tasks.
Each task can be handled only by a subset of the servers, requires a different level of
service, and once assigned can not be reassigned. We make the natural assumption
that the level of service is known at arrival time, but that the duration of service
is not. The online load balancing problem is to assign each task to an appropriate
server in such a way that the maximum load on the servers is minimized. In this
paper we derive matching upper and lower bounds for the competitive ratio of the
online greedy algorithm for this problem, namely (3n) 2=3
(1 + o(1)), and derive a
bound,\Omega\Gamma n 1=2 ), for any other deterministic or randomized online algorithm.
Consider an idealized local area network that links multimedia workstations, com
puters, I/O devices, etc. Each device is directly connected to one or more gateways
(bridges) to the net. Communication tasks arrive and disappear at arbitrary times.
Upon arrival, each task requests a certain guaranteed bandwidth (e.g. low for file